Abstract
Reporters obtained the following quote from the background information supplied by the inventors:“Advances in machine learning are providing new and unique opp ortunities for computers to aid humansby performing or automating tasks. For ex ample, a computer may be used to develop a predictive model.In supervised learn ing, a predictive model is developed through training. Training typically involv esestablishing and adapting a learning model (such as one or more neural networ ks, linear regression systems,logistic regression systems, Bayesian analysis sy stems, etc.) based on observations, such as examples of human behaviors, e.g., r ecognizing faces, identifying images, understanding language, performing routinedecisions, etc. Therefore, in order to perform the training, human action(s) us ually need to be well known,and the results of the action(s) need to be appropr iately labeled or classified, so that the parameters inthe predictive model can be determined. Moreover, training typically requires that a large number of examples or instances of the observations be measured or assembled.
NETL
NSTL